© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 4-RES-SM
    11. 14482
    Prev
    Next
    INDEX
    Explanations

    parentheses and other syntactical structures typically found in programming code

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/4-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.4.hook_resid_post
    Hook Layer
    4
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    78
    +0.17
    1.0%
    397
    +0.15
    0.8%
    131
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    397
    +0.17
    0.07
    131
    +0.15
    0.07
    315
    +0.12
    0.07
    Negative Logits
    ftware
    -1.47
    poons
    -1.42
     erupted
    -1.35
    ized
    -1.35
     abroad
    -1.30
    ization
    -1.30
     installation
    -1.28
    aneously
    -1.26
     correctly
    -1.26
     successfully
    -1.24
    POSITIVE LOGITS
    IJ
    4.06
    ĨĴ
    4.06
    ¡
    4.03
    ¬
    4.00
    Ģ
    3.99
    ¥
    3.97
    ¸
    3.93
    ī
    3.85
    ł
    3.79
    ŀ
    3.73
    Activations Density 0.402%

    No Known Activations